Cornelius Carl

@TweetsOfSumit @SysLambdaOmega Sollte eigentlich reichen für 5% (bei einigen MacBooks Pro Jahr)

@SysLambdaOmega maybe 20 macbooks, couple of phones, etc isn't "too many" yet

Curious if Apple ever wonders what tf is up with my private Apple Store Account...

< > < > > > 😵‍💫

@SadlyItsBradley they probably have some game-api specific way to be done, im just curious as usually things like this are pretty well done built in. maybe try an iphone app that you has haptics and see that one on visionOS? thx for your service

@ccoryoso I haven’t felt any system vibrations yet. I know the operating system supports that it for apps my controllers are like at 5% battery so maybe that’s a reason Or they need to actually be called by the app itself

PSVR2 controller is tracking and interacting with visionOS 26

The illusion of free choice

@AlexAlspaugh @doji_com looks so nice! currently does not let me continue with a german phone number as its too short for the continue button to be enabled. phone number length is 10 in Germany :)

But for now download @doji_com and let us know what you think apps.apple.com/us/app/doji-tr…

Doji is live now, this is the most fun I’ve ever had designing something and I hope everyone has even more fun using it

@theo @HiteshRohira15 @convex why not like the notion team using a sqlite database in OPFS? notion.com/blog/how-we-sp…

@HiteshRohira15 @convex IndexedDB is not stable enough to be truly local first in a web-focused application. I have a lot to share about the hell we've been through. The sync part was always more important. Adding a lot of local caching to keep the snappiness without the compromises :)

Been hard at work on rebuilding T3 Chat's sync layer. Works fully across multiple instances. Even across browsers and devices. Handles unstable network connections gracefully. This is gonna be so good. Coming soon™️

@eikedrescher HTML characters don't get properly decoded on the preview page, & stays &amp; :)

If I'm building a podcast app, maybe I should also have a podcast? "Some View" – Irregular short voice memos about designing apps like Bento & Queue, thinking about design & art, and other related things. Subscribe here: tryqueue.app/podcast?feed=h…

@newsystems_ brampton

It's finally here: Brampton Brampton is the world's most intelligent, creative, and fastest model. Brampton dramatically outperforms Grok 3, Claude 3.7 Sonnet, and GPT 4.5. Reply with "brampton" for early access.

@Q22635587 @Andercot @zoodotdev Old c based engine, pretty good tho. Grasshopper and rhino itself are hold back by leaving them in an early 2000s setting, eg C# plugin infrastructure and old ux. They could enable a lot of innovation by porting to web based but are hesitant bc legacy

@Andercot @zoodotdev What about grasshopper and Rhino???

Former professional CAD engineer weighing in on this. This is 100% the future of design. AI has 10-100x'd the productivity of software development but mechanical design is brutally tedious, slow, painful. @zoodotdev can do this for hardware. In an ideal world the major CAD design environments would be natively built around top-down parametric design when you can adjust a few key parameters and the entire build regenerates. Unfortunately SolidWorks is Just. Not. That. It's based on individual Parts; each part lives in its own file. OnShape *tried* to be this but its a tinker-toy when it comes to designing Large Serious Things that have hundreds of parts. It grinds to a halt. Every SolidWorks engineer knows the agony and pain of building a large complicated assembly and then having part references start breaking and suddenly everything cascades into a brutal mess and you're stuck manually re-defining hundreds of mating relationships and fixing broken part references. If you want to be part of a design team then you have to 'check out' parts one a time which basically locks the assembly if there are contextually defined features that depend on that part. Where this is really going is this: you want to be able to design an assembly of interrelated parts based on a few key driving parameters, and then pipe that mechanical design into an FEM or CFD simulations environment and look at a few key metrics. For fusion devices this was usually things like temperatures, gradients, given a certain applied thermal flux on a boundary condition, or else mechanical stresses e.g. produced by large superconducting magnets. You then take those simulation results and start adjusting those design parameters to get the performance metrics within a certain spec - max stress, max temperature, whatever. Now, OnShape at least saw the future and built in a scripting language to let people semi-automate their workflows, but there was a massive disconnect between designing things in the GUI and the scripting language itself. FeatureScript tbh just sorta sucked to work with, and they somehow missed the extremely obvious boat of getting an LLM to just generate FeatureScript based on plain text descriptions. The scripting back-end was just such a bespoke thing you had to reaaally love spending time on OnShape forums to get anything going. Even more important is that the future of CAD design starts to look a lot more like programming if you want to make hardware development as fluid and fast as software development at least within the design-simulate-refine cycle prior to fabricating anything. There's a very strong correspondence between principles of good Object Oriented Programming design and CAD design. There's many equivalent ways of getting the same function behavior or solid-body geometry but not all those ways are easily maintanable, adjustable, modified, etc. It's also rare that CAD designers also have OOP experience and so frequently the CAD design might look beautiful as a finished set of mechanical drawings but can become a nightmare to maintain itself as a feature tree of geometric operations. I think fundamentally Zoo is approaching all these things in the right way. The GUI Design environment is directly convertable to programming scripts right in the same environment. Every GUI action generates a line of code. You can see exactly how to modify the code to e.g. parameterize things, extend things, etc. Then, having an LLM that can turbo-charge the scripting process is a huge win. To do all these things they had to build an entirely new geometry kernel from the ground up, which is the thing that handles basic geometry operations like extrusions, lofts, intersections, etc etc. This is very non-trivial - if you dig under the hood of most any CAD design environment or even COMSOL you'll find the Parasolid Kernel, an ancient behemoth thats decades old and written in C code. Try scripting some basic operations directly in the Parasolid kernel, its pain. Its literally dozens of lines of code to do something like "make a round hole in this block" etc. nightmare. Basically if you enable this kind of top-down parametric design, with direct conversion between GUI actions and scripting interface, then add an LLM on top of that which can take up tedious operations like "add 12 x 14/-20 tapped holes half an inch from the edge of this mounting plate" etc etc - you can probably collapse the design time of a lot of hardware products by a factor of 10. People don't realize but CAD just hasn't had the same groundswell of constantly self-improving tooling as software devs for the simple reason that software devs build their own tools, but CAD designers don't write software and just have to deal with the shittiness of SolidWorks, CATIA, etc. The real win is if @zoodotdev can start to take on some of the simulations work as well all within the same environment either by making use of publicly available FEA or CFD codes or just implementing their own. Imagine text-to-cad making the design process itself 10x faster, and then when you want to simulate performance, instead of exporting some BREP object like a STEP file into COMSOL and applying a bunch of boundary conditions, visualizing the results, going back to CAD env and tuning things, re-exporting, etc etc etc - which is a process usually split across different people emailing back and forth slide decks and files - you just define the boundary conditions within your CAD environment and do something like "change these three parameters until internal stresses in this part get below XX GPa" and let the simulation run, modifying the CAD design directly with each iteration, propagating those changes to all the other contextually defined parts, until you get the thing you need. From design concept to fab-ready drawings that's probably a speed up more like 50x than 10x.

@maxim_perumal EU never disallowed, Apple did this in a petty move. All Intelligence features are coming in April. And the borders between countries in Europe are similar to borders of states in the US, so why have blocking in one of them. Hate on the EU for lack of startup investments not this

@TweetsOfSumit digitales wählen ist tatsächlich schwierig, cooles Video zu dem Thema: youtu.be/w3_0x6oaDmI?si…

Deutschland 🇩🇪 hat nicht genug Papier und Drucker für Neuwahlen aber warum kann ich denn nicht einfach faxen? Alternativ kann sich doch auch jede Straße einen Zettel teilen überlegt mal wie viele Kreuze auf ein DIN A4 Papier passen! WIR BRAUCHEN JETZT PRAGMATISCHE LÖSUNGEN!

@aboodman @thekitze @natebirdman @benjiDotSo @one__js this sounds like a perfect solution to the problem we are currently facing. Would love the possibility of early access

@thekitze @natebirdman @benjiDotSo @one__js You should try Zero. It's perfect for this kind of thing. I've sent you a DM.

ok i've fkn had it with react-query optimistic updates for @benjiDotSo... the todos and planner *need* to be local-first and sync without queries the app is made with blitz.js (next 15, postgresql, prisma) what's the suggested path? electricsql?

@jsngr how to solve distribution? web is king here + platform agnostic

i’ve been spending a lot of time with SwiftUI lately the next breakthroughs in AI interfaces will be made here

@jamiegraydesign @GimbalNick Do you have an example for that? Infinitely collapsible directed graphs sound so cool

@GimbalNick why not an infinitely collapsible directed graph why does everyone use the bad and dumb canvas idea I hate this timeline so much

@Azadux Hover over the address bar (where it stays the folder name), wait until the folder icon shows up and drag the icon into save dialog

ok macOS nerds - help me with this basic OS question I want to save a text file into a folder that I have open. What's the easiest way to copy the folder's address and paste it into the save file dialogue? I dont wanna navigate through folders in the save file dialogue--I have the folder open, I should be able to copy its directory and paste it as the save file location. in windows, this is arbitrary as the folder addresses are always exposed. In MacOS, I've found that I have to right click on the folder to copy the pathname, but how do i paste it into the save file dialogue?

@carmguti @thiamineenjoyer pubmed.ncbi.nlm.nih.gov/19695149/ I think it’s the combination of both

@thiamineenjoyer im pretty sure the cancer association was from the repeated drinking of boiling water, not mate

@daniel_friis Big fan of your work, do you by any chance have the urbanvillageproject website archived? Keep coming back for the ux&story telling :)

Late last year, my family and I made a big change to our lives. We moved from the comfort of our home city in Copenhagen to a new life in a small village in Portugal. We have been longing to try something new for some time now and with the closing of SPACE10 and my wife as an entrepreneur, it was finally possible. We settled on Portugal for many reasons. First and foremost because of the culture. It feels like people have a different pace to life than we're used to. Slower, but in a positive way where life and people around you are absorbed and appreciated in a different way. The sunny weather is also a welcome improvement from Denmark. To be honest, it's been more work than we had anticipated. And leaving your comfort zone is always, well, uncomfortable. But the upsides vastly outpace the drawbacks. It's been especially rewarding to witness our children adjusting to a new environment. Children are so adaptable and resilient, and we trust that an experience like this will help build their confidence and perspective on other cultures. In terms of myself and this new chapter after SPACE10, I’ve decided to go back to my craft and dive into the familiar world of code and the new world of AI. I'm intrigued by what the new technologies have to offer and what I believe it can do for human potential. I'm currently working on a few experiments which I find really interesting, including one aimed at enabling more accessible mental health support, one in children's entertainment and empowerment, and one in AI-assisted writing. Over the next few months, I'll be sharing some of these and more under the umbrella of Nowadays (nowadays.tech) to get some feedback and learnings. I'm very curious to hear what you think and will take any input and feedback I can get 🙏 Aside from this, I will also be doing a bit of advisory work and investing. If any of the above resonate with you, I'd love to talk! :) From sunny Portugal ☀️ Daniel

@carmguti lmao

@herval he recently mentioned it paying for the avp price in the first 24 hours, so I would assume sales for the first year to be <100k

As a counterpoint to this: the guy who developed Apollo for Reddit probably bagged close to a million dollars already, just from day-one sales of his YouTube wrapper for AVP